Modellar va narxlar

1M TOKEN UCHUN

API'ning har bir model id'si, u nimani qabul qilishi va qancha turishi: 10 ta Shannon modeli va 12 ta xostingdagi open-weight model, har biri uchala endpointda mavjud. GET /v1/models xuddi shu ro'yxatni JSON sifatida qaytaradi.

senzurasiz model: 22
kontekst oynasi: 256K
dan: $0.50 /1M

Xostingdagi open-weight modellar

Har bir id'da 262,144 tokenli kontekst; kirish va chiqish alohida hisoblanadi; har bir id o'zi taqlid qilgan model qabul qiladigan narsani qabul qiladi

Model	Kirish	Tuzilmali chiqish	1M uchun Kirish / kesh / chiqish	Ko'proq
DeepSeek-V4-Pro-0813`3BIT-REAP`DeepSeek · 1.6T MoE · 49B activeSenzurasiz	Matn kirishi	JSON sxemasi	$1.95 / $0.488 / $3.90	Ko'proq
GLM-5.2`3BIT-REAP`Z.ai · 744B MoE · 40B activeSenzurasiz	Matn kirishi	JSON sxemasi	$0.73 / $0.183 / $2.34	Ko'proq
Kimi-K3`3BIT-REAP`Moonshot AI · 2.8T MoE · 104B activeSenzurasiz	Matn + rasm kirishi	JSON sxemasi	$3.83 / $0.958 / $19.12	Ko'proq
Nemotron3Ultra`3BIT-REAP`NVIDIA · 550B hybrid Mamba-Attention MoE · 55B activeSenzurasiz	Matn kirishi	JSON sxemasi	$0.75 / $0.188 / $3.30	Ko'proq
MiniMax-M3`3BIT-REAP`MiniMax · 428B MoE · 23B active · sparse attentionSenzurasiz	Matn + rasm kirishi	JSON sxemasi	$0.50 / $0.125 / $2.00	Ko'proq
DeepSeek-V4-Flash-0731`W4A16-AUTOROUND-REAP`DeepSeek · 284B MoE · 13B activeSenzurasiz	Matn kirishi	JSON sxemasi	$0.50 / $0.125 / $2.00	Ko'proq
Kimi-K2.6`W4A16-AUTOROUND-REAP`Moonshot AI · 1T MoE · 32B activeSenzurasiz	Matn + rasm kirishi	JSON sxemasi	$0.78 / $0.195 / $3.67	Ko'proq
Laguna-S-2.1`W4A16-AUTOROUND-REAP`Poolside · 118B MoE · 8B activeSenzurasiz	Matn kirishi	Tuzilmali chiqish yo'q	$0.50 / $0.125 / $2.00	Ko'proq
inkling`W4A16-AUTOROUND-REAP`Thinking Machines · 975B MoE · 41B activeSenzurasiz	Matn + rasm kirishi	JSON obyekti	$1.42 / $0.355 / $6.07	Ko'proq
MiMo-V2.5-Pro`W8A16`Xiaomi · 1.02T MoE · 42B active · 8-bitSenzurasiz	Matn kirishi	JSON sxemasi	$0.50 / $0.125 / $2.00	Ko'proq
MiMo-V2.5`W8A16`Xiaomi · 8-bit multimodal · W8A16Senzurasiz	Matn + rasm kirishi	JSON sxemasi	$0.50 / $0.125 / $2.00	Ko'proq
Hy3`W8A16`Tencent · 295B MoE · 21B active · 8-bitSenzurasiz	Matn kirishi	JSON sxemasi	$0.50 / $0.125 / $2.00	Ko'proq

Shannon modellari

kirish va chiqish uchun bitta qat'iy narx

Model	Kontekst	Kirish	Tuzilmali chiqish	Narx / 1M token
Shannon 1.6 Lite`shannon-1.6-lite`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$3.90
Shannon 1.6 Pro`shannon-1.6-pro`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$7.80
Shannon 2 Lite`shannon-2-lite`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$3.90
Shannon 2 Pro`shannon-2-pro`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$5.85
Shannon 3`shannon-3`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$3.35
Shannon 3 Pro`shannon-3-pro`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$3.35
Shannon 3.1`shannon-3.1`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$3.35
Shannon 3.1 Pro`shannon-3.1-pro`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$3.35
Shannon 3.5`shannon-3.5`Senzurasiz	192K	Matn + rasm kirishi	JSON sxemasi	$3.35
Shannon Coder 1`shannon-coder-1`Senzurasiz	128K	Matn kirishi	JSON sxemasi	$8.00

Tez boshlashga o'tish Playgroundni sinab ko‘ring

Qatorni qanday o'qish kerak

Ustun	Nimani bildiradi
Model	Ko'rsatiladigan nom. Uning ostidagi qator `model` sifatida yuboradigan id. Xostingdagi jadvalda id — nom va uning yonidagi teg, defis bilan birlashtirilgan: `MiniMax-M3` va `3BIT-REAP` `MiniMax-M3-3BIT-REAP`ni hosil qiladi. Nom ostidagi qator modelni kim yaratganini va uning o'lchamini ko'rsatadi.
Kontekst	Kontekst oynasi: model suhbatning nechta tokenini o'qiydi. K — 1,024 token. Aniq son `GET /v1/models` dagi `context_window`.
Kirish	Faqat matn yoki matn va rasmlar. Rasmlarni o'qiydigan model `capabilities.vision: true`ni ko'rsatadi. `shannon-2-lite` va `shannon-2-pro` `tools` yoki `response_format` ham yuboradigan so'rovlarda rasmlarni o'qiydi.
Tuzilmali chiqish	`response_format` modeldan nimani so'rashi mumkin: JSON sxemangizga amal qiladigan javob, JSON obyekti bo'lgan javob yoki hech biri.
Narx	1,000,000 token uchun USD. Shannon modelida kirish va chiqish uchun bitta narx. Xostingdagi open-weight modelda uchta bor: kirish, keshdagi kirish va chiqish.

Har bir model streaming qiladi va tool chaqira oladi. API'ning boshqa imkoniyatlari Imkoniyatlar sahifasida tasvirlangan. Imkoniyatlar

Kontekst oynasi va chiqish limitlari

Kontekst oynasi model bitta javob uchun o'qiydigan hamma narsani qamrab oladi: xabarlaringiz, tool ta'riflari va rasmlar.
Shannon 3 oilasi uzunroq suhbatni o'zi sig'diradi: oynaga sig'adigan eng yangi navbatlarni saqlaydi va javob beradi.
Oqimda juda uzun prompt invalid_request_error turi va code: "context_length_exceeded" bilan xato freymi sifatida bildiriladi.
max_tokens har bir modelda 1 dan 65,536 gacha qiymatlarni qabul qiladi; GET /v1/models dagi max_output_tokens shu oraliqning yuqori chegarasi. Standart 4,096.
Har bir modelda max_tokens so'rov balansingizdan ajratib qo'yadigan miqdor.
Javob uzunligi limiti sifatida max_tokensni xostingdagi open-weight modellar, shannon-1.6-lite, shannon-1.6-pro va shannon-coder-1 qo'llaydi.
Xostingdagi open-weight modellarda u javob matnini cheklaydi. Javobdan oldingi mulohaza unga hisoblanmaydi, 256 dan past qiymat esa 256 kabi ishlaydi.
shannon-1.6-lite, shannon-1.6-pro, shannon-coder-1 yoki Shannon 3 oilasi oqimi javob limitida uzilganda finish_reason: "length" bilan tugaydi.

Limitlar va balans

Narx qanday shakllanadi

Narxlar 1,000,000 token uchun USD da. Balansingiz 1,000,000 uchun $5.00 turadigan tokenlarda sanaladi, shuning uchun model narxi uning tokenlari balansdan qancha olishini aytadi.
Shannon modellarining bitta narxi bor. Chaqiruv shu narxda usage.total_tokens bo'yicha hisoblanadi.
Xostingdagi open-weight modellarning uchta narxi bor. Prompt keshidan o'qilgan kirish keshdagi narxda hisoblanadi, u kirish narxining 25% i, $0.001 gacha yaxlitlangan. Kirishning qolgan qismi kirish narxida, chiqish chiqish narxida hisoblanadi.
Xostingdagi open-weight modellar tokenlarni har bir modelning o'z tokenizatori bilan sanaydi. GET /v1/models buni tokenizer: "original" deb belgilaydi.
shannon-coder-1 /v1/chat/completions va /v1/messages da tokenda emas, chaqiruvda hisoblanadi. Uning ko'rsatilgan narxi /v1/responses da qo'llanadi.

So'rov nimani ajratib qo'yishi, hisoblangan misollar va Shannon Coder chaqiruv limiti Limitlar va balans sahifasida. Limitlar va balans

Model id'lari

model sifatida shu id'lardan birini yuboring.

Model id'lari

shannon-1.6-lite
shannon-1.6-pro
shannon-2-lite
shannon-2-pro
shannon-3
shannon-3-pro
shannon-3.1
shannon-3.1-pro
shannon-3.5
shannon-coder-1
DeepSeek-V4-Pro-0813-3BIT-REAP
GLM-5.2-3BIT-REAP
Kimi-K3-3BIT-REAP
Nemotron3Ultra-3BIT-REAP
MiniMax-M3-3BIT-REAP
DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP
Kimi-K2.6-W4A16-AUTOROUND-REAP
Laguna-S-2.1-W4A16-AUTOROUND-REAP
inkling-W4A16-AUTOROUND-REAP
MiMo-V2.5-Pro-W8A16
MiMo-V2.5-W8A16
Hy3-W8A16

Katta va kichik harflar ahamiyatsiz, id atrofidagi bo'sh joylar olib tashlanadi: SHANNON-3 va shannon-3 bir xil model.
Qaysi yozuvni yuborganingizdan qat'i nazar, javob modelda doim e'lon qilingan id'ni olib keladi.
Shannon 3 kabi ko'rsatiladigan nom id emas.

To'rtta qisqaroq yozuv ham qabul qilinadi.

Siz yuborasiz	Javob beradigan model
`shannon-1.6`	`shannon-1.6-lite`
`shannon-3-lite`	`shannon-3`
`shannon-3.1-lite`	`shannon-3.1`
`shannon-coder`	`shannon-coder-1`

E'lon qilinmagan id'ga uchala endpointda ham 400 holati, invalid_request_error turi va unknown model: <id> xabari bilan javob beriladi. Uning o'rniga boshqa model javob bermaydi.

{
  "error": {
    "type": "invalid_request_error",
    "message": "unknown model: gpt-4o"
  }
}

{
  "type": "error",
  "error": {
    "type": "invalid_request_error",
    "message": "unknown model: gpt-4o"
  }
}

modelni doim yuboring. /v1/responses usiz so'rovga 400 bilan javob beradi. /v1/chat/completions va /v1/messages da bunday so'rovga shannon-1.6-lite javob beradi.
POST /v1/tokenize va POST /v1/messages/count_tokens xostingdagi open-weight id'larni qabul qiladi.

Modellar ro'yxati

GET /v1/models shu sahifadagi modellarni xuddi shu kontekst oynalari, narxlar va imkoniyatlar bilan JSON sifatida qaytaradi. Undan model tanlagichni to'ldirish yoki kodda narxni tekshirish uchun foydalaning.

GET https://api.shannon-ai.com/v1/models

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.shannon-ai.com/v1"
)

for model in client.models.list():
    print(model.id)

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'YOUR_API_KEY',
  baseURL: 'https://api.shannon-ai.com/v1'
});

for await (const model of client.models.list()) {
  console.log(model.id);
}

curl "https://api.shannon-ai.com/v1/models"

Javob, har bir turdan bitta yozuvgacha qisqartirilgan:

200 JSON

{
  "object": "list",
  "data": [
    {
      "id": "shannon-3",
      "object": "model",
      "created": 1700000000,
      "owned_by": "shannon-ai",
      "display_name": "Shannon 3",
      "context_window": 196608,
      "max_output_tokens": 65536,
      "pricing": {
        "input_per_million_usd": 3.35,
        "output_per_million_usd": 3.35,
        "currency": "USD"
      },
      "capabilities": {
        "tools": true,
        "vision": true,
        "reasoning": true,
        "reasoning_effort": false,
        "streaming": true,
        "json_schema": true
      },
      "endpoints": [
        "/v1/chat/completions",
        "/v1/messages",
        "/v1/responses"
      ]
    },
    {
      "id": "MiniMax-M3-3BIT-REAP",
      "object": "model",
      "created": 1700000000,
      "owned_by": "shannon-ai",
      "display_name": "MiniMax-M3 · 3BIT-REAP",
      "context_window": 262144,
      "max_output_tokens": 65536,
      "pricing": {
        "input_per_million_usd": 0.5,
        "cached_input_per_million_usd": 0.125,
        "output_per_million_usd": 2,
        "currency": "USD"
      },
      "tokenizer": "original",
      "capabilities": {
        "tools": true,
        "vision": true,
        "reasoning": true,
        "reasoning_effort": true,
        "streaming": true,
        "response_format": true,
        "json_schema": true,
        "prompt_caching": true
      },
      "endpoints": [
        "/v1/chat/completions",
        "/v1/messages",
        "/v1/responses",
        "/v1/messages/count_tokens",
        "/v1/tokenize"
      ]
    }
  ]
}

Maydon	Tur	Tavsif	Mavjud
`object`	`string`	Doim `list`.
`data`	`array`	Chaqirish mumkin bo'lgan har bir modelga bitta yozuv.
`data[].id`	`string`	`model` sifatida yuboradigan id.	Har bir model
`data[].object`	`string`	Doim `model`.	Har bir model
`data[].created`	`integer`	Qat'iy son, har bir model uchun bir xil: `1700000000`.	Har bir model
`data[].owned_by`	`string`	Doim `shannon-ai`.	Har bir model
`data[].display_name`	`string`	Ko'rsatish uchun nom. U `model` sifatida qabul qilinmaydi.	Har bir model
`data[].context_window`	`integer`	Kontekst oynasi, token bilan.	Har bir model
`data[].max_output_tokens`	`integer`	So'rov so'rashi mumkin bo'lgan eng katta chiqish byudjeti: har bir modelda `65536`.	Har bir model
`data[].pricing.input_per_million_usd`	`number`	1,000,000 kirish tokeni narxi, USD da.	Har bir model
`data[].pricing.cached_input_per_million_usd`	`number`	Prompt keshidan o'qilgan 1,000,000 kirish tokeni narxi.	Xostingdagi open-weight modellar
`data[].pricing.output_per_million_usd`	`number`	1,000,000 chiqish tokeni narxi, USD da. Shannon modelida u kirish narxiga teng.	Har bir model
`data[].pricing.currency`	`string`	Doim `USD`.	Har bir model
`data[].tokenizer`	`string`	`original`: tokenlar modelning o'z tokenizatori bilan sanaladi.	Xostingdagi open-weight modellar
`data[].capabilities.tools`	`boolean`	Model siz e'lon qilgan tool'larni chaqira oladi. Har bir modelda `true`.	Har bir model
`data[].capabilities.vision`	`boolean`	Model rasmlarni o'qiydi.	Har bir model
`data[].capabilities.reasoning`	`boolean`	Model javob berishdan oldin mulohaza qiladi.	Har bir model
`data[].capabilities.reasoning_effort`	`boolean`	`reasoning_effort` so'rov maydoni bu modelda kuchga kiradi.	Har bir model
`data[].capabilities.streaming`	`boolean`	Har bir modelda `true`.	Har bir model
`data[].capabilities.json_schema`	`boolean`	`json_schema` turidagi `response_format` kuchga kiradi.	Har bir model
`data[].capabilities.response_format`	`boolean`	`response_format` umuman, hech bo'lmasa JSON obyekti sifatida kuchga kiradi.	Xostingdagi open-weight modellar
`data[].capabilities.prompt_caching`	`boolean`	Takrorlanuvchi prompt prefikslari keshlanadi va keshdagi narxda hisoblanadi.	Xostingdagi open-weight modellar
`data[].endpoints`	`array`	Bu id'ni qabul qiladigan yo'llar.	Har bir model

GET /v1/models kalit talab qilmaydi. Javob har bir chaqiruvchi uchun bir xil va rejangizga bog'liq emas.
Ro'yxat chaqirish mumkin bo'lgan id'larni o'z ichiga oladi. Hozir xizmat ko'rsata olmaydigan model undan chiqarib tashlanadi.
Bitta model uchun endpoint yo'q: GET /v1/models/{id} not_found_error turi bilan 404 bilan javob beradi.